Scalable Event Stream Processing of Trigger Rules

This U.S. non-provisional application claims priority to U.S. provisional application No. 63/720,899, entitled “SCALABLE STREAM PROCESSING OF TRIGGER RULES,” and filed Nov. 15, 2024, the entirety of which is incorporated herein by reference.

Trigger rules are automated conditions set to activate specific actions when certain criteria are met. According to trigger rules, events and/or data are monitored to detect when predefined conditions are satisfied in order to initiate an automated response, such as, but not limited to, sending an alert, updating data, invoking a function, and/or executing a process. Trigger rules help automate workflows, and/or react to specific conditions to make processes more efficient by reducing the need for manual intervention. Trigger conditions can be customized to ensure that tasks, alerts, and/or processes are triggered under certain predefined conditions.

This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.

Systems, methods, and computer program products are disclosed for scalable event stream processing of trigger rules. Event data associated with a set of object is partitioned into a set of logical partitions. The set of logical partitions are provided to a set of event processors to enable the set of event processors to process event data in the logical partitions according to a set of trigger rules. In embodiments, the event processors cause actions to be performed in response to the satisfaction of the trigger rules.

Further features and advantages of the embodiments, as well as the structure and operation of various embodiments, are described in detail below with reference to the accompanying drawings. It is noted that the claimed subject matter is not limited to the specific embodiments described herein. Such embodiments are presented herein for illustrative purposes only. Additional embodiments will be apparent to persons skilled in the relevant art(s) based on the teachings contained herein.

The subject matter of the present application will now be described with reference to the accompanying drawings. In the drawings, like reference numbers indicate identical or functionally similar elements. Additionally, the left-most digit(s) of a reference number identifies the drawing in which the reference number first appears.

The following detailed description discloses numerous example embodiments. The

scope of the present patent application is not limited to the disclosed embodiments, but also encompasses combinations of the disclosed embodiments, as well as modifications to the disclosed embodiments. It is noted that any section/subsection headings provided herein are not intended to be limiting. Embodiments are described throughout this document, and any type of embodiment may be included under any section/subsection. Furthermore, embodiments disclosed in any section/subsection may be combined with any other embodiments described in the same section/subsection and/or a different section/subsection in any manner.

Trigger rules are automated conditions set to activate specific actions when certain criteria are met. According to trigger rules, events and/or data are monitored to detect when predefined conditions are satisfied. Trigger rules are often employed in event processing to determine when event data satisfies the trigger condition. Event processing includes simple event processing, where individual events are handled independently, and complex event processing, where multiple events are analyzed to identify patterns, correlations, and/or trends. Additionally, event processing include batch processing, where events are analyzed in batches, and stream processing, where event streams are analyzed in real time or near-real time. As the size and/or complexity of monitored systems increase, the amount of data and/or events that need to be analyzed can exceed the capacity of a single machine. In such scenarios, it becomes necessary to partition the event data in order to distribute event processing across a plurality of machines. This is especially true for real-time event stream processing, where delayed processing of event data can result in a delayed response that may result in undesirable results (e.g., security breaches, QoS violations, etc.).

Existing techniques for partitioning event data include key-based partitioning, where event data is partitioned based on a key (e.g., identifier) in the event data, and the partitioned event data is assigned to a task for processing based on a hash of the key. This approach enables event data that shares a common key (e.g., identifier) to be processed together by the task to which the partition is assigned. However, key-based partitioning of event data can result in a large number of partitions, which can lead to substantial overhead costs. Furthermore, complex event processing can include trigger rules that rely on event data from different event data streams associated with different identifiers. In such scenarios, key-based partitioning lacks the flexibility to route event data from multiple event data streams to the same event processor for processing. For instance, when event data relied upon by a trigger rule are routed to different event processors, the event processors may need to communicate with each other to determine when the trigger condition is satisfied, resulting in increased latency in detecting when the trigger condition is satisfied.

Embodiments disclosed herein overcome these issues in conventional systems. In order to reduce overhead costs associated with a large number of partitions, examples disclosed herein determine a number of logical partitions necessary to ensure that the event data is distributed across a sufficient number of event processors to prevent overloading, and assign a subset of the logical partitions to a customer (e.g., tenant) for processing event data belonging to the customer. In order to reduce latency in detecting when the trigger condition is satisfied, rule hints are employed that are determined from the trigger rules to route event data that is relevant to a trigger rule to the same partition for processing by the same event processor.

According to embodiments, a number of logical partitions for partitioning the event data is determined to ensure that the event data is distributed across a number of event processors to prevent overloading. For instance, a data partitioner determines a number of event processors needed to process the event data based on the capacity of the event processors, and determines the number of logical partitions based on the needed number of event processors. The data partitioner then assigns a subset of the logical partitions to a particular customer (e.g., tenant), and partitions event data belonging to the customer into the logical partitions assigned to the customer. The number of logical partitions assigned to the customer is limited based on the amount of event data belonging to the customer in order to reduce overhead.

The data partitioner, in embodiments, partitions the event data into a set of logical partitions using on rule hints determined from trigger rules, which allows a set of event processors to efficiently process the event data in a scalable fashion. For instance, trigger rules are analyzed to determine information (e.g., object identifiers) relevant (e.g., referenced by) the trigger rules, and rule hints are generated based on the determined information. Based on the rule hints, the data partitioner partitions event data associated with the determined information (e.g., object identifiers) into the same logical partition to ensure that the event processor that will process the partition has access to the event data necessary for determining whether the trigger rule is satisfied.

In configurations, event data arrives at an ingestion pipeline as event data associated with (e.g., generated by) one or more monitored objects, such as, but not limited to, entities (e.g., persons, organizations, etc.), items (e.g., packages, vehicles, components, etc.), locations (e.g., buildings, regions, etc.), and/or the like. The event data arrives in one or more formats, such as, but not limited to, as one or more data streams, in data batches, as time-series data, as timestamped data, as vector data, and/or the like. In embodiments, the event data includes information associated with one or more objects (e.g., entity, item, location, etc.) and/or object instances, such as, but not limited to, temperature, speed, time, utilization, sensor data, and/or the like.

The ingestion pipeline, in configurations, processes the event data against stateless trigger rules to determine whether the event data satisfies the stateless trigger rules. Employing an ingestion pipeline to process stateless trigger rules reduces the latency for triggering stateless rules. Additionally, stateless trigger rule processing is offloaded from the event processors, thereby freeing up additional capacity for the event processors to process complex trigger rules. In embodiments, the stateless trigger rules comprising stateless conditions that can be satisfied by one or more time-series values in the event data without maintaining state information for the time-series values.

The event data is, in configurations, provided to a data partitioner for partitioning into a set of logical partitions according to various factors, such as, but not limited to, the relevancy of event data to one or more of a set of trigger rules, the capacity of the set of event processors, and/or the like. As used herein, the term “logical partition” refers to an abstract subset of event data that will processed by the same event processor. For instance, event data is logically divided based on rule hints into subsets of event data and the subsets of event data are provided to event processors for processing. Partitioning event data based on rule hints reduces latency in detecting when a trigger condition is satisfied by ensuring that event data relevant to a trigger rule are partitioned into the same logical partition for processing by the same the event processor. In embodiments, event data associated with a logical partition can be repartitioned by dividing the logical partition into two or more logical partitions, and/or reassigned to a different event processor for processing.

In examples, the set of trigger rules are analyzed to determine a set of rule hints that are used to partition the event data. As used herein, the term “rule hint” refers to any information relevant to the trigger rules, such as, but not limited to, the types of objects (e.g., entity type, item type, location type, etc.) relevant to the trigger rules, and/or the types of information (e.g., object identifier, temperature, speed, time, etc.) relevant to the trigger rules. Based on the rule hints, the data partitioner partitions the event data into a set of logical partitions such that event data relevant to the same trigger rule is grouped together. In embodiments, the data is partitioned according to the type of object (e.g., entity type, item type, location type, etc.) such that event data related to different types of objects are partitioned into different logical partitions. In instances, the data is partitioned according to object instance such that event data related to different instances of the same type of object are partitioned into different logical partitions, and/or partitioned according to the type of information (e.g., temperature, speed, time, etc.) such that event data related to different types of information are partitioned into different logical partitions.

The data partitioner, in configurations, assigns available partitions to customers (e.g., tenants) based on an identifier associated with the customer (e.g., tenantID). In instances, the data partitioner employs a rendezvous hashing algorithm to determine which partitions to assign to which customers. In embodiments, event data associated with (e.g., belonging to) the customer is split into the partitions assigned to the customer using a second hashing algorithm based on an instance identifier associated with the object instance. Partitioning event data into the partitions assigned to the customer reduces the overhead costs associated with existing key-based partitioning techniques by employing fewer partitions.

In embodiments, the data partitioner partitions the event data according to the capacities of the set of event processors in order to ensure that the logical partitions can be processed by a single event processor. For instance, the data partitioner determines the throughput of the event processors (e.g., number of events processed per time period) and determines the logical partitions to ensure that the number of events per time period in the logical partitions does not exceed the throughput capacity of the event processors. In embodiments, the data partitioner determines updated capacities of the set of event processors and repartitions the event data using the updated capacities. In instances, the throughput capacity of the event processors are limited by the available resources (e.g., processor capacity, memory capacity, etc.) of the event processors.

The data partitioner, in configurations, provides the set of logical partitions to the set of event processors using the capacities of the set of event processors. For instance, the data partitioner groups one or more logical partitions using the size of the logical partitions and/or the capacities of the set of event processors, and provides the groups of logical partitions to the set of event processors. In embodiments, an event processor of the set of event processors process one or more of the logical partitions to determine whether event data in the one or more logical partitions satisfy trigger conditions associated with a subset of the trigger rules. By partitioning the event data into a set of logical partitions, the set of event processors can process large amounts of event data in a scalable and efficient fashion.

In aspects, the data partitioner stores the set of logical partitions in a storage system for future access. For instance, when a trigger rule is added and/or activated, the stored set of logical partitions can be replayed by providing the stored logical partitions to an event processor to enable the event processor to determine whether a newly added and/or activated trigger rule is satisfied by past event data. In embodiments, the stored set of logical partitions are stored in various ways, such as, but not limited to, in a database, in a file storage system, in a cloud storage system, as a blob (Binary Large Object), and/or the like.

The set of event processors, in examples, acquire ownership of available (i.e., unclaimed) logical partitions in order to process event data associated with the available partitions. In instances, when an event processors gains ownership of a partition, it performs a recovery process to load state data in order to start processing event associated with the partition, and when an event processor loses ownership of a partition, it stops processing of event data associated with the partition. The ability of event processors to acquire and lose ownership of a logical partition enhances load balancing by redistributing logical partitions from overloaded event processors to other event processors, and failover by enabling logical partitions owned by a failed event processor to be transferred to another event processor.

The set of event processors, in configurations, process event data in the set of logical partitions in order to determine whether the event data satisfy trigger conditions defined by the set of trigger rules. For instance, the event processors maintain state information associated with the objects using the event data in the set of logical partitions, and determine whether the state information satisfies the trigger conditions defined by the set of trigger rules.

In embodiments, the trigger rules comprise a set of one or more functional expressions that are evaluated over time-series data. For example, a trigger rule for determining whether a temperature exceeds a threshold temperature (e.g., 32 degrees F, etc.) will include a functional expression that compares temperature values (e.g., 10 F, 20 F, 30 F, 34 F, 43 F, etc.) in the time-series data against the threshold temperature and outputs a series of Boolean values (e.g., 0, 0, 0, 1, 1, etc.) that are indicative of whether the temperature values of the time-series data exceed the threshold temperature. In such a scenario, the trigger rule would be satisfied when the output includes a Boolean value equal to TRUE (e.g., 1). In another example, a trigger rule for determining whether a temperature remains over a threshold temperature (e.g., 32 degrees F, etc.) for a predetermined period of time (e.g., 10 minutes) will include a first functional expression that compares temperature values (e.g., 10 F, 20 F, 30 F, 34 F, 43 F, etc.) in the time-series data against the threshold temperature and outputs a series of Boolean values (e.g., 0, 0, 0, 1, 1, etc.) that are indicative of whether the temperature values of the time-series data exceed the threshold temperature, and a second functional expression that counts the number of output Boolean values that are equal to TRUE (e.g., 1) to determine whether the temperature has remained over the threshold temperature for the predetermined time period. In embodiments, trigger rules can combine any number of functional expressions, such as, but not limited to, logical expressions (e.g., <, <=, =, >=, >, etc.), mathematical expressions (e.g., average, mean, median, percentage, count, standard deviation, etc.), sliding window expressions (e.g., count over a sliding window, mathematical function over a sliding window, etc.), and/or the like. In embodiments, the event processors maintain state information associated with the objects by storing one or more outputs of the functional expressions.

When an event processor determines that a trigger condition is satisfied, the event processor, in examples, causes a responsive action to be performed, such as, but not limited to, transmitting a message and/or alert to a recipient, opening a ticket in a ticketing system, invoking a webhook, invoking an API (Application Programming Interface) function, automatically performing a remedial action, and/or the like. In embodiments, the responsive action is determined by a user (e.g., customer, admin, etc.) when defining the trigger rule.

In order to enable load balancing and/or failover, the set of event processors, in configurations, stores the state information associated with objects in a storage system for access by other event processors. In embodiments, the set of event processors periodically store a progress indicator in a storage system to indicate the event data that has been processed by the event processor. In examples, the state information and/or the progress indicators are stored in various ways, such as, but not limited to, in a database, in a file storage system, in a cloud storage system, as a blob (Binary Large Object), and/or the like.

In order to provide load balancing and/or failover, the set of logical partitions can, in embodiments, be redistributed to different event processors for processor, and/or repartitioned into a different number of logical partitions to balance the load across the event processors. For instance, when a first event processor becomes unavailable (e.g., fails, requires maintenance, becomes unstable, etc.), or becomes overloaded (e.g., has high utilization, has low capacity, etc.) the data partitioner transfers processing of one or more logical partitions being processed by the first event processor to second event processor. The second event processor accesses progress indicators and state information associated with the transferred logical partitions, and processes event data in the transferred logical partitions starting from a position determined using the progress indicators.

In embodiments, the data partitioner increases the number of partitions assigned to the customer in order to spread out the data processing across additional event processors. In examples, the data partitioner increases the number of partitions assigned to a customer by a factor of two (i.e., x2) in order to simplify checkpoint management. For instance, by doubling the number of partitions by a factor of 2, event data belonging to each partition is divided between two new (i.e., child) partitions, and the new partitions can refer to the checkpoints of the old (i.e., parent) partition during checkpoint management. In configurations, the data partitioner repartitions the data when the number of trigger rules and/or when the number of partitions assigned to a customer changes. For instance, the data partitioner determines updated partitioning parameters and an effective time for the updated partitioning, and provides the updated partitioning parameters and effective time to the event processors. In instances, the data partitioner determines the effective time as an arbitrary time in the future that provides sufficient notice to the event processors to transition to processing of the event data based on the updated partitioning parameters. At the effective time, the data partitioner begins partitioning the event data based on the updated partitioning parameters, and the event processors stop processing event data that no longer flow to the partitions associated with the event processor and initiate a recovery process to load state data in order to start processing event data for data newly flowing to the partitions associated with the event processor.

The amount of event data in the set of logical partitions can change over time based on the type of event data in incoming event data streams. For instance, the amount of event data generated by an object instance can fluctuate over time, resulting in a change in the amount of data in a logical partition that includes event data from the object instance. These fluctuations can cause the amount of event data in a logical partition to exceed a processing capacity of an event processor and/or cause an imbalance in the distribution of the set of logical partitions among the set of event processors. In such scenarios, the data partitioner, in embodiments, repartitions the event data into a set of updated logical partitions that are distributed to the set of logical partitions. In order to adjust the number of event processors according to the workload, one or more event processors is added and/or removed from the set of event processors responsive to the increase and/or decrease in the amount of event data. Adjusting the number of event processors optimizes resource utilization by balancing performance (e.g., satisfying throughput/latency goals) and cost (e.g., avoiding underutilized event processors). After acquiring ownership of updated logical partitions, the set of event processors, in examples, access progress indicators and/or state information associated with the updated logical partitions, and process event data in the updated logical partitions starting from a position determined using the progress indicators.

In embodiments, the addition and/or activation of a new trigger rule to the set of trigger rules requires the processing of past event data to determine whether the new trigger rule is satisfied. In such instances, the event processor replays event data relevant to the new trigger rule by accessing past event data from storage, and updating state information associated with an object using the past event data to determine whether the state information satisfies a trigger condition associated with the new trigger rule. In examples, the event processor replays the event data from a point in time necessary to determine whether the trigger condition has been satisfied. For instance, the event processor replays event data in the last 24 hours when a new trigger rule has a trigger condition that is dependent on 24 hours of event data.

These and further embodiments enable the functionality described above and additional functionality. Such embodiments are described in further detail as follows.

1 FIG. 1 FIG. 100 100 102 104 106 108 110 100 For example,shows a block diagram of an example systemfor scalable event stream processing of trigger rules, in accordance with an embodiment. As shown in, systemincludes a server infrastructurethat comprises an ingestion pipeline, a data partitioner, rules data, and one or more event processors. Systemis described in further detail as follows.

102 102 102 1170 11 FIG. Server infrastructurecomprises a network-accessible server set (e.g., cloud-based environment or platform). In an embodiment, the underlying resources of server infrastructureare co-located (e.g., housed in one or more nearby buildings with associated components such as backup power supplies, redundant data communications, environmental controls, etc.) to form a datacenter, are distributed across different regions, and/or are arranged in other manners. Various example implementations of server infrastructureare described below in reference to(e.g., network-based server infrastructure, and/or components thereof).

104 112 104 112 112 106 114 104 2 FIG. Ingestion pipelineis configured to receive event datafrom a customer. In embodiments, ingestion pipelinereceives event datafrom a customer through various means, such as, but not limited to, directly from a customer, through a monitoring agent executing on a component associated with the customer, from storage associated with the customer, and/or the like. In embodiments, ingestion engine provides at least a portion of event datato data partitioneras event datafor partitioning into a set of logical partitions. Ingestion pipelinewill be described in greater detail below in conjunction with.

112 112 In embodiments, event datais associated with one or more monitored objects and/or object instances, such as, but not limited to, entities (e.g., persons, organizations, etc.), items (e.g., packages, vehicles, components, etc.), locations (e.g., buildings, regions, etc.), and/or the like. In embodiments, event dataarrives in various formats, such as, but not limited to, as one or more data streams, in data batches, as time-series data, as timestamped data, as vector data, and/or the like. In embodiments, the event data includes information associated with one or more objects (e.g., entity, item, location, etc.) and/or object instances, such as, but not limited to, temperature, speed, time, utilization, sensor data, and/or the like.

106 114 118 106 116 116 116 106 114 118 106 106 106 114 106 114 106 114 106 118 110 110 106 118 118 110 118 110 110 106 Data partitioneris configured to partition event datainto a set of logical partitionsaccording to various factors, such as, but not limited to, the relevancy of event data to one or more of a set of trigger rules, the capacity of the set of event processors, and/or the like. In embodiments, data partitioneremploys a set of rule hintsto partition the event data. For instance, rule hintscan specify the information relevant to the trigger rules, such as, but not limited to, the types of objects (e.g., entity type, item type, location type, etc.) relevant to the trigger rules, and/or the types of information (e.g., temperature, speed, time, etc.) relevant to the trigger rules. Using rule hints, data partitioner, in embodiments, partitions event datainto logical partitionssuch that relevant event data are grouped together. In instances, data partitioneremploys a rendezvous hashing algorithm to determine which partitions to assign to which customers. In embodiments, data partitionerpartitions event data associated with the customer into the partitions assigned to the customer using a second hashing algorithm based on an instance identifier associated with the object instance. In embodiments, data partitionerpartitions event dataaccording to the type of object (e.g., entity type, item type, location type, etc.) such that event data related to different types of objects are partitioned into different logical partitions. In embodiments, data partitionerpartitions event dataaccording to object instance such that event data related to different instances of the same type of object are partitioned into different logical partitions. In embodiments, data partitionerpartitions event dataaccording to the type of information (e.g., temperature, speed, time, etc.) such that event data related to different types of information are partitioned into different logical partitions. In embodiments, data partitionerprovides logical partitionsto event processor(s)according to the capacities of event processor(s). For instance, data partitionergroups one or more logical partitionsaccording to the size of logical partitionsand/or the capacities of event processor(s), and provides the groups of logical partitionsto event processor(s)according to the capacities of event processor(s). In embodiments, data partitionerassigns available partitions to customers (e.g., tenants) based on an identifier associated with the customer (e.g., tenantID).

106 114 106 118 110 106 110 106 110 114 118 106 110 114 118 110 204 114 110 106 2 3 FIGS.and In embodiments, data partitionerrepartitions the event datawhen the number of trigger rules and/or when the number of partitions assigned to a customer changes. For instance, data partitionerincreases the number of logical partitionsassigned to the customer in order to spread out the data processing across additional event processor(s). In embodiments, data partitionerdetermines updated partitioning parameters and an effective time for the updated partitioning, and provides the updated partitioning parameters and effective time to event processor(s). In embodiments, data partitionerdetermines the effective time as an arbitrary time in the future that provides sufficient notice to event processor(s)to transition to processing of event databased on updated logical partitions. In embodiments, data partitionerbegins partitioning the event data based on the updated partitioning parameters at the effective time. In embodiments, at the effective time, event processor(s)stop processing event datathat no longer flow to logical partitionsassociated with event processor(s), and initiate a recovery process to load state datain order to start processing event datafor data newly flowing to the partitions associated with event processor(s). Data partitionerwill be described in greater detail below in conjunction with.

108 120 120 Rules datacomprise trigger rulesthat include one or more trigger conditions and one or more responsive actions. In embodiments, the trigger conditions comprise a set of one or more functional expressions that are evaluated over time-series data. In embodiments, trigger rulescombine any number of functional expressions, such as, but not limited to, logical expressions (e.g., <, <=, =, >=, >, etc.), mathematical expressions (e.g., average, mean, median, percentage, count, standard deviation, etc.), sliding window expressions (e.g., count over a sliding window, mathematical function over a sliding window, etc.), and/or the like.

110 118 118 120 122 120 110 118 114 118 110 110 110 110 110 1174 1146 1102 1192 110 11 FIG. 2 4 FIGS.- Event processor(s)comprise computational nodes configured to process logical partitionsto determine whether event data in logical partitionssatisfy trigger conditions of trigger rules, and to perform responsive actionsof trigger rulesresponsive to the satisfaction of the trigger conditions. In embodiments, event processor(s)acquire ownership of available (i.e., unclaimed) logical partitionsin order to process event dataassociated with logical partitions. In embodiments, when event processor(s)gain ownership of a partition, event processor(s)perform a recovery process to load state data in order to start processing event associated with the partition. In embodiments, when event processor(s)loses ownership of a partition, event processor(s)stop processing of event data associated with the partition. Various example implementations of event processor(s)are described below in reference to(e.g., nodes, node, computing device, on-premises servers, and/or components thereof). Event processor(s)will be described in greater detail below in conjunction with.

2 FIG. 2 FIG. 200 200 102 104 106 108 110 200 102 202 204 206 104 208 110 210 212 214 200 Embodiments described herein may operate in various ways to perform scalable event stream processing of a logical partition by an event processor. For instance,shows a block diagram of an example systemfor scalable event stream processing of a logical partition by an event processor, in accordance with an embodiment. As shown in, systemcomprises server infrastructure, ingestion pipeline, data partitioner, rules data, and an event processorA. In system, server infrastructurefurther includes event data, state data, and progress data. Furthermore, ingestion pipelinefurther includes a stateless rule executer, and event processorA further includes a state updater, a stateful rule executer, and a progress updater. Systemis described in further detail as follows.

202 220 118 106 118 202 202 Event datais configured to store event dataassociated with logical partitions. In embodiments, data partitionerstores logical partitions, and/or event data thereof, in event datafor future access. In embodiments, event datais implemented in various ways, such as, but not limited to, as a database, as a file storage system, as a cloud storage system, as blob storage, and/or the like.

204 222 204 State datais configured to store stateassociated with one or more monitored objects. In embodiments, state datais implemented in various ways, such as, but not limited to, as a database, as a file storage system, as a cloud storage system, as blob storage, and/or the like.

206 224 118 224 118 110 206 Progress datais configured to store progress indicatorsassociated with the processing of logical partitions. In embodiments, progress indicatorsindicate the event data of logical partitionsthat have been processed by event processor(s). In embodiments, progress datais implemented in various ways, such as, but not limited to, as a database, as a file storage system, as a cloud storage system, as blob storage, and/or the like.

208 112 112 216 218 216 208 112 Stateless rule executeris configured to process event datato determine whether event data in event datasatisfies a trigger condition of a stateless trigger rule, and to cause a responsive actionto be performed responsive to the satisfaction of the trigger condition of stateless trigger rule. For example, stateless rule executercan determine whether a stateless trigger rule for determining whether a temperature exceeds a threshold temperature (e.g., 32 degrees F, etc.) is satisfied by comparing temperature values (e.g., 10 F, 20 F, 30 F, 34 F, 43 F, etc.) in event dataagainst the threshold temperature without maintaining any state information.

110 110 110 118 118 120 216 112 Event processorA is an event processor of event processor(s). In embodiments, event processorA is configured to receive a subset of logical partitionsand process event data in the subset of logical partitionsto determine whether trigger conditions of a subset of trigger rulesare satisfied. In embodiments, stateless trigger rulecomprises a stateless condition that can be satisfied by event datawithout maintaining state information.

210 118 106 222 210 118 210 222 212 204 State updateris configured to receive logical partitionsfrom data partitionerand update a stateassociated with a monitored object. For instance, state updatedcan update a cumulative value (e.g., counter, average, average over a sliding time window, etc.) using incoming event data in logical partitions. In embodiments, state updaterprovides updated stateto stateful rule executerand/or state data.

212 222 210 120 108 222 120 212 222 120 212 122 Stateful rule executeris configured to receive statefrom state updaterand a trigger rulefrom rules data, and to determine whether statesatisfies a trigger condition associated with trigger rule. In embodiments, when stateful rule executerdetermines that statesatisfies the trigger condition associated with trigger rule, stateful rule executercauses responsive actionto be performed, such as, but not limited to, transmitting a message and/or alert to a recipient, opening a ticket in a ticketing system, invoking a webhook, invoking an API (Application Programming Interface) function, automatically performing a remedial action, and/or the like. In embodiments, the responsive action is determined by a user (e.g., customer, admin, etc.) when defining the trigger rule.

214 224 206 214 224 206 214 224 Progress updateris configured to store progress indicatorin progress data. In embodiments, progress updaterstores progress indicatorin progress dataat regular time intervals. In embodiments, progress updaterstores progress indicatorat specific milestones, such as, but not limited to, the satisfaction of a trigger conditions, and/or the like.

3 FIG. 3 FIG. 300 300 102 104 106 108 110 204 206 210 212 214 300 Embodiments described herein may operate in various ways to transfer event stream processing of a logical partition to another event processor. For instance,shows a block diagram of an example systemfor transferring event stream processing of a logical partition to another event processor, in accordance with an embodiment. As shown in, systemcomprises server infrastructure, ingestion pipeline, data partitioner, rules data, an event processorB, state data, progress data, state updater, stateful rule executer, and progress updater. Systemis described in further detail as follows.

110 118 110 110 110 106 118 110 110 210 110 304 302 118 118 304 Event processorB is a transfer destination for a logical partitionthat is being transferred from another event processor(e.g., event processorA). For instance, when an event processor (e.g., event processorA) becomes unavailable (e.g., fails, requires maintenance, becomes unstable, etc.), or becomes overloaded (e.g., has high utilization, has low capacity, etc.) data partitionertransfers processing of one or more logical partitionsbeing processed by the event processor (e.g., event processorA) to another event processor (e.g., event processorB). In embodiments, state updaterof event processorB accesses progress indicatorand stateassociated with a transferred logical partition, and processes event data in the transferred logical partitionstarting from a position determined using progress indicator.

4 FIG. 4 FIG. 400 400 102 108 110 202 204 206 210 212 214 400 Embodiments described herein may operate in various ways to perform event stream processing of a new trigger rule. For instance,shows a block diagram of an example systemfor event stream processing of a new trigger rule, in accordance with an embodiment. As shown in, systemcomprises server infrastructure, rules data, an event processorC, event data, state data, progress data, state updater, stateful rule executer, and progress updater. Systemis described in further detail as follows.

110 110 404 404 402 Event processorC is an event processor of event processor(s)configured to process past event datato determine whether past event datasatisfies a trigger condition associated with new trigger rule.

402 404 402 210 110 404 202 222 404 212 110 222 402 406 402 222 402 210 110 404 402 In embodiments, new trigger rulecomprises a newly added and/or activated trigger rule that requires the processing of past event datato determine whether a trigger condition associated with new trigger ruleis satisfied. In embodiments, state updaterof event processorC accesses past event datafrom event data, and updates stateusing past event data. In embodiments, stateful rule executerof event processorC determines whether statesatisfies a trigger condition associated with new trigger rule, and causes a responsive actionof new trigger ruleto be performed responsive to determining that statesatisfies a trigger condition associated with new trigger rule. In embodiments, state updaterevent processorC accesses past event datafrom a position necessary to determine whether the trigger condition of new trigger rulehas been satisfied.

5 FIG. 1 2 FIGS.and 500 102 500 500 Embodiments described herein may operate in various ways to perform scalable event stream processing of logical partition by an event processor. For instance,depicts a flowchartof a process for scalable event stream processing of logical partition by an event processor, in accordance with an embodiment. Server infrastructure, may, for example, operate according to flowchart. Flowchartis described as follows with respect tofor illustrative purposes.

500 502 502 104 112 104 112 112 106 114 Flowchartstarts at step. In step, event data associated with a set of objects is received. For example, ingestion pipelinereceives event data. In embodiments, ingestion pipelinereceives event datafrom a customer through various means, such as, but not limited to, directly from a customer, through a monitoring agent executing on a component associated with the customer, from storage associated with the customer, and/or the like. In embodiments, ingestion engine provides at least a portion of event datato data partitioneras event datafor partitioning into a set of logical partitions.

504 106 116 118 114 106 114 106 114 106 118 110 110 106 118 118 110 118 110 110 In step, a set of logical partitions for partitioning the event data is determined based on rule hints determined from a set of event trigger rules, a first logical partition of the set of logical partitions comprising first event data associated with a first object of the set of objects. For example, data partitionerdetermines, based on rule hints, a set of logical partitionsfor partitioning event data. In embodiments, data partitionerpartitions dataaccording to object instance such that event data related to different instances of the same type of object are partitioned into different logical partitions. In embodiments, data partitionerpartitions event dataaccording to the type of information (e.g., temperature, speed, time, etc.) such that event data related to different types of information are partitioned into different logical partitions. In embodiments, data partitionerprovides logical partitionsto event processor(s)according to the capacities of event processor(s). For instance, data partitionergroups one or more logical partitionsaccording to the size of logical partitionsand/or the capacities of event processor(s), and provides the groups of logical partitionsto event processor(s)according to the capacities of event processor(s).

506 106 118 110 106 118 118 110 118 110 210 222 118 210 118 In step, the first logical partition is provided to a first event processor of a set of event processors to enable the first event processor to update a state associated with the first object using the first event data in the first logical partition. For example, data partitionerprovides logical partitionsto event processor(s). In embodiments, data partitionergroups one or more logical partitionsaccording to the size of the logical partitionsand/or the capacities of event processor(s), and provides the groups of logical partitionsto event processor(s)according to the capacities. In embodiments, state updaterupdates stateusing event data in logical partitions. For instance, state updatedcan update a cumulative value (e.g., counter, average, average over a sliding time window, etc.) using incoming event data in logical partitions.

508 212 222 120 In step, the state associated with the first object is determined to satisfy a first trigger condition associated with a first event trigger rule of the set of event trigger rules. For example, stateful rule executerdetermines that statesatisfies a trigger condition associated with trigger rule.

510 212 122 In step, a first action associated with the first event trigger rule is performed. For example, stateful rule executorcauses responsive actionto be performed, such as, but not limited to, transmitting a message and/or alert to a recipient, opening a ticket in a ticketing system, invoking a webhook, invoking an API (Application Programming Interface) function, automatically performing a remedial action, and/or the like.

6 FIG. 1 2 FIGS.and 600 102 600 600 Embodiments described herein may operate in various ways to partition event data into a set of logical partitions. For instance,depicts a flowchartof a process for partitioning event data into a set of logical partitions, in accordance with an embodiment. Server infrastructure, may, for example, operate according to flowchart. Flowchartis described as follows with respect tofor illustrative purposes.

600 602 602 106 110 Flowchartstarts at step. In step, a set of capacities associated with a set of event processors is determined. For example, data partitionerdetermines a set of capacities associated with event processor(s).

604 106 118 114 110 106 118 110 114 In step, the set of logical partitions is determined based on the set of capacities associated with the event processors. For example, data partitionerdetermines logical partitionsfor partitioning event datausing capacities associated with event processor(s). In embodiments, data partitionerdetermines logical partitionsbased on the number of event processor(s)needed to process event data.

7 FIG. 1 3 FIGS.- 700 102 700 700 Embodiments described herein may operate in various ways to repartition event data into a set of logical partitions. For instance,depicts a flowchartof a process for repartitioning event data into a set of logical partitions, in accordance with an embodiment. Server infrastructure, may, for example, operate according to flowchart. Flowchartis described as follows with respect tofor illustrative purposes.

700 702 702 106 110 Flowchartstarts at step. In step, a set of updated capacities associated with a set of event processors is determined. For example, data partitionerdetermines a set of updated capacities associated with event processor(s).

704 106 118 114 110 118 110 In step, a set of updated logical partitions for repartitioning the event data is determined based on the set of updated capacities, a first updated logical partition of the set of updated logical partitions comprising second event data associated with the first object. For example, data partitionerdetermines updated logical partitionsfor repartitioning event databased on capacities associated with event processor(s), and provides updated logical partitionsto event processor(s).

706 210 110 302 204 In step, state information associated with the first object is accessed by a second event processor of a set of event processor. For example, state updaterof event processorB accesses statefrom state data.

708 118 210 110 210 110 302 118 In step, the first updated logical partition is provided to the second event processor to enable the second event processor to update a state associated with the first object using the second event data in the first updated logical partition. For example, data partitioner provides updated logical partitionto state updaterof event processorB, and state updaterof event processorB updates stateusing event data in updated logical partition.

8 FIG. 1 3 FIGS.- 800 102 800 800 800 800 Embodiments described herein may operate in various ways to transfer event stream processing of a logical partition to another event processor. For instance,depicts a flowchartof a process for transferring event stream processing of a logical partition to another event processor, in accordance with an embodiment. Server infrastructure, may, for example, operate according to flowchart. Note that not all steps of flowchartneed to be performed in all embodiments, and in some embodiments, the steps of flowchartmay be performed in different orders than shown. Flowchartis described as follows with respect tofor illustrative purposes.

800 802 802 214 110 224 206 224 118 110 214 224 206 214 224 Flowchartstarts at step. In step, a progress indicator is periodically stored by a first event processor, the progress indicator indicative of first event data processed by the first event processor. For example, progress updaterof event processorA periodically stores progress indicatorsin progress data. In embodiments, progress indicatorsindicate the event data of logical partitionsthat have been processed by event processor(s). In embodiments, progress updaterstores progress indicatorin progress dataat regular time intervals. In embodiments, progress updaterstores progress indicatorat specific milestones, such as, but not limited to, the satisfaction of a trigger conditions, and/or the like.

804 106 118 110 110 In step, processing of a first logical partition is transferred from the first event processor to a second event processor of a set of event processors. For example, data partitionertransfers processing of logical partitionfrom event processorA to event processorB.

806 210 110 304 302 118 118 304 In step, the progress indicator and state information associated with the first object is accessed by the second event processor. For example, state updaterof event processorB accesses progress indicatorand stateassociated with a transferred logical partition, and processes event data in the transferred logical partitionstarting from a position determined using progress indicator.

808 106 118 110 110 In step, the first logical partition is redirected from the first event processor to the second event processor. For example, data partitionerredirects logical partitionfrom event processorA to event processorB.

810 210 110 118 304 302 In step, the first event data in the first logical partition is processed by the second event processor starting from a position determined using the progress indicator. For example, state updaterof event processorB processes event data in logical partitionstarting from a position determined using progress indicatorin order to update state.

9 FIG. 1 2 4 FIGS.,, and 900 102 900 900 900 900 Embodiments described herein may operate in various ways to perform event stream processing of a new trigger rule. For instance,depicts a flowchartof a process for event stream processing of a new trigger rule, in accordance with an embodiment. Server infrastructure, may, for example, operate according to flowchart. Note that not all steps of flowchartneed to be performed in all embodiments, and in some embodiments, the steps of flowchartmay be performed in different orders than shown. Flowchartis described as follows with respect tofor illustrative purposes.

900 902 902 402 108 Flowchartstarts at step. In step, a new event trigger rule is added to a set of event trigger rules, the new event trigger rule associated with a second object of a set of objects. For example, new trigger ruleis added to and/or activated in rules data.

904 210 110 404 202 In step, event data associated with the second object is replayed. For example, state updaterof event processorC accesses past event datafrom event data.

906 210 110 222 404 In step, a state associated with the second object is updated using the event data associated with the second object. For example, state updaterof event processorC updates stateusing past event data.

908 212 110 222 402 In step, the state associated with the second object is determined to satisfy a second trigger condition associated with the new event trigger rule. For example, stateful rule executerof event processorC determines that statesatisfies a trigger condition of new trigger rule.

910 212 110 406 In step, a second action associated with the new event trigger rule is performed. For example, stateful rule executerof event processorC causes responsive actionto be performed.

10 FIG. 1 2 FIGS.and 1000 102 1000 1000 Embodiments described herein may operate in various ways to perform event stream processing of a stateless event trigger rule. For instance,depicts a flowchartof a process for event stream processing of a stateless event trigger rule, in accordance with an embodiment. Server infrastructure, may, for example, operate according to flowchart. Flowchartis described as follows with respect tofor illustrative purposes.

1000 1002 1002 208 112 216 Flowchartstarts at step. In step, prior to partitioning event data, it is determined that the event data satisfies a second trigger condition associated with a stateless event trigger rule. For example, stateless rule executerdetermines that event data of event datasatisfies a trigger condition of a stateless trigger rule.

1004 208 218 In step, a second action associated with the stateless event trigger rule is performed. For example, stateless rule executercauses a responsive actionto be performed.

102 104 106 108 110 202 204 206 208 210 212 214 500 600 700 800 900 1000 104 106 108 110 202 204 206 208 210 212 214 500 600 700 800 900 1000 102 104 106 108 110 202 204 206 208 210 212 214 500 600 700 800 900 1000 Server infrastructure, ingestion pipeline, data partitioner, rules data, event processor(s), event data, state data, progress data, stateless rule executer, state updater, stateful rule executer, progress updater, and/or the components described therein, and/or the steps of flowcharts,,,,, and/orare implemented in hardware, or hardware combined with one or both of software and/or firmware. For example, ingestion pipeline, data partitioner, rules data, event processor(s), event data, state data, progress data, stateless rule executer, state updater, stateful rule executer, progress updater, and/or the components described therein, and/or the steps of flowcharts,,,,, and/orare each implemented as computer program code/instructions configured to be executed in one or more processors and stored in a computer readable storage medium. Alternatively, server infrastructure, ingestion pipeline, data partitioner, rules data, event processor(s), event data, state data, progress data, stateless rule executer, state updater, stateful rule executer, progress updater, and/or the components described therein, and/or the steps of flowcharts,,,,, and/orare implemented in one or more SoCs (system on chip). An SoC includes an integrated circuit chip that includes one or more of a processor (e.g., a central processing unit (CPU), microcontroller, microprocessor, digital signal processor (DSP), etc.), memory, one or more communication interfaces, and/or further circuits, and optionally executes received program code and/or include embedded firmware to perform functions.

11 FIG. 11 FIG. 11 FIG. 1100 1102 1102 102 110 1102 1102 1100 1104 1104 1104 1104 1102 Embodiments disclosed herein can be implemented in one or more computing devices that are mobile (a mobile device) and/or stationary (a stationary device) and include any combination of the features of such mobile and stationary computing devices. Examples of computing devices in which embodiments are implementable are described as follows with respect to.shows a block diagram of an exemplary computing environmentthat includes a computing device. Computing deviceis an example of server infrastructure, event processor(s), and/or components described therein, which each include one or more of the components of computing device. In some embodiments, computing deviceis communicatively coupled with devices (not shown in) external to computing environmentvia network. Networkcomprises one or more networks such as local area networks (LANs), wide area networks (WANs), enterprise networks, the Internet, etc. In examples, networkincludes one or more wired and/or wireless portions. In some examples, networkadditionally or alternatively includes a cellular network for cellular communications. Computing deviceis described in detail as follows.

1102 1102 1102 Computing devicecan be any of a variety of types of computing devices. Examples of computing deviceinclude a mobile computing device such as a handheld computer (e.g., a personal digital assistant (PDA)), a laptop computer, a tablet computer, a hybrid device, a notebook computer, a netbook, a mobile phone (e.g., a cell phone, a smart phone, etc.), a wearable computing device (e.g., a head-mounted augmented reality and/or virtual reality device including smart glasses), or other type of mobile computing device. In an alternative example, computing deviceis a stationary computing device such as a desktop computer, a personal computer (PC), a stationary server device, a minicomputer, a mainframe, a supercomputer, etc.

11 FIG. 11 FIG. 1102 1110 1120 1142 1144 1130 1150 1160 1180 1182 1184 1186 1120 1156 1122 1124 1188 1120 1112 1114 1116 1160 1162 1164 1166 1150 1152 1154 1130 1132 1134 1136 1138 1140 1102 1102 1102 1102 1102 1102 As shown in, computing deviceincludes a variety of hardware and software components, including a processor, a storage, a graphics processing unit (GPU), a neural processing unit (NPU), one or more input devices, one or more output devices, one or more wireless modems, one or more wired interfaces, a power supply, a location information (LI) receiver, and an accelerometer. Storageincludes memory, which includes non-removable memoryand removable memory, and a storage device. Storagealso stores an operating system, application programs, and application data. Wireless modem(s)include a Wi-Fi modem, a Bluetooth modem, and a cellular modem. Output device(s)includes a speakerand a display. Input device(s)includes a touch screen, a microphone, a camera, a physical keyboard, and a trackball. Not all components of computing deviceshown inare present in all embodiments, additional components not shown may be present, and in a particular embodiment any combination of the components are present. In examples, components of computing deviceare mounted to a circuit card (e.g., a motherboard) of computing device, integrated in a housing of computing device, or otherwise included in computing device. The components of computing deviceare described as follows.

1110 1110 1102 1110 1110 1112 1114 1120 1110 1112 1102 1114 1114 1110 1144 1142 In embodiments, a single processor(e.g., central processing unit (CPU), microcontroller, a microprocessor, signal processor, ASIC (application specific integrated circuit), and/or other physical hardware processor circuit) or multiple processorsare present in computing devicefor performing such tasks as program execution, signal coding, data processing, input/output processing, power control, and/or other functions. In examples, processoris a single-core or multi-core processor, and each processor core is single-threaded or multithreaded (to provide multiple threads of execution concurrently). Processoris configured to execute program code stored in a computer readable medium, such as program code of operating systemand application programsstored in storage. The program code is executable to cause processorto perform operations, including the processes/methods disclosed herein. Operating systemcontrols the allocation and usage of the components of computing deviceand provides support for one or more application programs(also referred to as “applications” or “apps”). In examples, application programsinclude common computing applications (e.g., e-mail applications, calendars, contact managers, web browsers, messaging applications), further computing applications (e.g., word processing applications, mapping applications, media player applications, productivity suite applications), one or more machine learning (ML) models, as well as applications related to the embodiments disclosed elsewhere herein. In examples, processor(s)includes one or more general processors (e.g., CPUs) configured with or coupled to one or more hardware accelerators, such as one or more NPUsand/or one or more GPUs.

1102 1106 1110 1102 1106 11 FIG. Any component in computing devicecan communicate with any other component according to function, although not all connections are shown for ease of illustration. For instance, as shown in, busis a multiple signal line communication medium (e.g., conductive traces in silicon, metal traces along a motherboard, wires, etc.) present to communicatively couple processorto various other components of computing device, although in other embodiments, an alternative bus, further buses, and/or one or more individual signal lines is/are present to communicatively couple components. Busrepresents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures.

1120 1156 1188 1112 1114 1116 1122 1122 1110 1122 1118 1118 1124 1102 1102 1124 1188 1102 1188 11 FIG. Storageis physical storage that includes one or both of memoryand storage device, which store operating system, application programs, and application dataaccording to any distribution. Non-removable memoryincludes one or more of RAM (random access memory), ROM (read only memory), flash memory, a solid-state drive (SSD), a hard disk drive (e.g., a disk drive for reading from and writing to a hard disk), and/or other physical memory device type. In examples, non-removable memoryincludes main memory and is separate from or fabricated in a same integrated circuit as processor. As shown in, non-removable memorystores firmwarethat is present to provide low-level control of hardware. Examples of firmwareinclude BIOS (Basic Input/Output System, such as on personal computers) and boot firmware (e.g., on smart phones). In examples, removable memoryis inserted into a receptacle of or is otherwise coupled to computing deviceand can be removed by a user from computing device. Removable memorycan include any suitable removable memory device type, including an SD (Secure Digital) card, a Subscriber Identity Module (SIM) card, which is well known in GSM (Global System for Mobile Communications) communication systems, and/or other removable physical memory device type. In examples, one or more of storage deviceare present that are internal and/or external to a housing of computing deviceand are or are not removable. Examples of storage deviceinclude a hard disk drive, a SSD, a thumb drive (e.g., a USB (Universal Serial Bus) flash drive), or other physical storage device.

1120 1112 1114 104 106 108 110 202 204 206 208 210 212 214 500 600 700 800 900 1000 One or more programs are stored in storage. Such programs include operating system, one or more application programs, and other program modules and program data. Examples of such application programs include computer program logic (e.g., computer program code/instructions) for implementing ingestion pipeline, data partitioner, rules data, event processor(s), event data, state data, progress data, stateless rule executer, state updater, stateful rule executer, progress updater, and/or each of the components described therein, as well as any of flowcharts,,,,,, and/or any individual steps thereof.

1120 1112 1114 1116 1116 1116 1120 Storagealso stores data used and/or generated by operating systemand application programsas application data. Examples of application datainclude web pages, text, images, tables, sound files, video data, and other data. In examples, application datais sent to and/or received from one or more network servers or other devices via one or more wired or wireless networks. Storagecan be used to store further data including a subscriber identifier, such as an International Mobile Subscriber Identity (IMSI), and an equipment identifier, such as an International Mobile Equipment Identifier (IMEI). Such identifiers can be transmitted to a network server to identify users and equipment.

1102 1130 1102 1150 1130 1132 1134 1136 1138 1140 1150 1152 1154 1130 1150 1102 1102 1102 1102 1180 1160 1130 1154 1132 1130 1150 1134 1136 1152 1154 In examples, a user enters commands and information into computing devicethrough one or more input devicesand receives information from computing devicethrough one or more output devices. Input device(s)includes one or more of touch screen, microphone, camera, physical keyboardand/or trackballand output device(s)includes one or more of speakerand display. Each of input device(s)and output device(s)are integral to computing device(e.g., built into a housing of computing device) or are external to computing device(e.g., communicatively coupled wired or wirelessly to computing devicevia wired interface(s)and/or wireless modem(s)). Further input devices(not shown) can include a Natural User Interface (NUI), a pointing device (computer mouse), a joystick, a video game controller, a scanner, a touch pad, a stylus pen, a voice recognition system to receive voice input, a gesture recognition system to receive gesture input, or the like. Other possible output devices (not shown) can include piezoelectric or other haptic output devices. Some devices can serve more than one input/output function. For instance, displaydisplays information, as well as operating as touch screenby receiving user commands and/or other information (e.g., by touch, finger gestures, virtual keyboard, etc.) as a user interface. Any number of each type of input device(s)and output device(s)are present, including multiple microphones, multiple cameras, multiple speakers, and/or multiple displays.

1142 1142 1142 In embodiments where GPUis present, GPUincludes hardware (e.g., one or more integrated circuit chips that implement one or more of processing cores, multiprocessors, compute units, etc.) configured to accelerate computer graphics (two-dimensional (2D) and/or three-dimensional (3D)), perform image processing, and/or execute further parallel processing applications (e.g., training of neural networks, etc.). Examples of GPUperform calculations related to 3D computer graphics, include 2D acceleration and framebuffer capabilities, accelerate memory-intensive work of texture mapping and rendering polygons, accelerate geometric calculations such as the rotation and translation of vertices into different coordinate systems, support programmable shaders that manipulate vertices and textures, perform oversampling and interpolation techniques to reduce aliasing, and/or support very high-precision color spaces.

1144 1128 1144 1144 In examples, NPU(also referred to as an “artificial intelligence (AI) accelerator” or “deep learning processor (DLP)”) is a processor or processing unit configured to accelerate artificial intelligence and machine learning applications, such as execution of machine learning (ML) model (MLM). In an example, NPUis configured for a data-driven parallel computing and is highly efficient at processing massive multimedia data such as videos and images and processing data for neural networks. NPUis configured for efficient handling of AI-related tasks, such as speech recognition, background blurring in video calls, photo or video editing processes like object detection, etc.

1144 1128 1128 In embodiments disclosed herein that implement ML models, NPUcan be utilized to execute such ML models, of which MLMis an example. For instance, where applicable, MLMis a generative AI model that generates content that is complex, coherent, and/or original. For instance, a generative AI model can create sophisticated sentences, lists, ranges, tables of data, images, essays, and/or the like. An example of a generative AI model is a language model. A language model is a model that estimates the probability of a token or sequence of tokens occurring in a longer sequence of tokens. In this context, a “token” is an atomic unit that the model is training on and making predictions on. Examples of a token include, but are not limited to, a word, a character (e.g., an alphanumeric character, a blank space, a symbol, etc.), a sub-word (e.g., a root word, a prefix, or a suffix). In other types of models (e.g., image based models) a token may represent another kind of atomic unit (e.g., a subset of an image). Examples of language models applicable to embodiments herein include large language models (LLMs), text-to-image AI image generation systems, text-to-video AI generation systems, etc. A large language model (LLM) is a language model that has a high number of model parameters. In examples, an LLM has millions, billions, trillions, or even greater numbers of model parameters. Model parameters of an LLM are the weights and biases the model learns during training. Some implementations of LLMs are transformer-based LLMs (e.g., the family of generative pre-trained transformer (GPT) models). A transformer is a neural network architecture that relies on self-attention mechanisms to transform a sequence of input embeddings into a sequence of output embeddings (e.g., without relying on convolutions or recurrent neural networks).

1144 1128 1128 1128 1128 1128 1128 1128 1128 1128 1144 1128 In further examples, NPUis used to train MLM. To train MLM, training data is that includes input features (attributes) and their corresponding output labels/target values (e.g., for supervised learning) is collected. A training algorithm is a computational procedure that is used so that MLMlearns from the training data. Parameters/weights are internal settings of MLMthat are adjusted during training by the training algorithm to reduce a difference between predictions by MLMand actual outcomes (e.g., output labels). In some examples, MLMis set with initial values for the parameters/weights. A loss function measures a dissimilarity between predictions by MLMand the target values, and the parameters/weights of MLMare adjusted to minimize the loss function. The parameters/weights are iteratively adjusted by an optimization technique, such as gradient descent. In this manner, MLMis generated through training by NPUto be used to generate inferences based on received input feature sets for particular applications. MLMis generated as a computer program or other type of algorithm configured to generate an output (e.g., a classification, a prediction/inference) based on received input features, and is stored in the form of a file or other data structure.

1128 1144 1128 1144 1128 In examples, such training of MLMby NPUis supervised or unsupervised. According to supervised learning, input objects (e.g., a vector of predictor variables) and a desired output value (e.g., a human-labeled supervisory signal) train MLM. The training data is processed, building a function that maps new data on expected output values. Example algorithms usable by NPUto perform supervised training of MLMin particular implementations include support-vector machines, linear regression, logistic regression, Naïve Bayes, linear discriminant analysis, decision trees, K-nearest neighbor algorithm, neural networks, and similarity learning.

1128 1128 In an example of supervised learning where MLMis an LLM, MLMcan be trained by exposing the LLM to (e.g., large amounts of) text (e.g., predetermined datasets, books, articles, text-based conversations, webpages, transcriptions, forum entries, and/or any other form of text and/or combinations thereof). In examples, training data is provided from a database, from the Internet, from a system, and/or the like. Furthermore, an LLM can be fine-tuned using Reinforcement Learning with Human Feedback (RLHF), where the LLM is provided the same input twice and provides two different outputs and a user ranks which output is preferred. In this context, the user's ranking is utilized to improve the model. Further still, in example embodiments, an LLM is trained to perform in various styles, e.g., as a completion model (a model that is provided a few words or tokens and generates words or tokens to follow the input), as a conversation model (a model that provides an answer or other type of response to a conversation-style prompt), as a combination of a completion and conversation model, or as another type of LLM model.

1128 1128 1128 1128 1128 1144 1128 According to unsupervised learning, MLMis trained to learn patterns from unlabeled data. For instance, in embodiments where MLMimplements unsupervised learning techniques, MLMidentifies one or more classifications or clusters to which an input belongs. During a training phase of MLMaccording to unsupervised learning, MLMtries to mimic the provided training data and uses the error in its mimicked output to correct itself (i.e., correct weights and biases). In further examples, NPUperform unsupervised training of MLMaccording to one or more alternative techniques, such as Hopfield learning rule, Boltzmann learning rule, Contrastive Divergence, Wake Sleep, Variational Inference, Maximum Likelihood, Maximum A Posteriori, Gibbs Sampling, and backpropagating reconstruction errors or hidden state reparameterizations.

1144 1110 1142 1144 1128 Note that NPUneed not necessarily be present in all ML model embodiments. In embodiments where ML models are present, any one or more of processor, GPU, and/or NPUcan be present to train and/or execute MLM.

1160 1102 1110 1102 1104 1160 1166 1160 1164 1162 1162 1164 One or more wireless modemscan be coupled to antenna(s) (not shown) of computing deviceand can support two-way communications between processorand devices external to computing devicethrough network, as would be understood to persons skilled in the relevant art(s). Wireless modemis shown generically and can include a cellular modemfor communicating with one or more cellular networks, such as a GSM network for data and voice communications within a single cellular network, between cellular networks, or between the mobile device and a public switched telephone network (PSTN). In examples, wireless modemalso or alternatively includes other radio-based modem types, such as a Bluetooth modem(also referred to as a “Bluetooth device”) and/or Wi-Fi modem(also referred to as an “wireless adaptor”). Wi-Fi modemis configured to communicate with an access point or other remote Wi-Fi-capable device according to one or more of the wireless network protocols based on the IEEE (Institute of Electrical and Electronics Engineers) 802.11 family of standards, commonly used for local area networking of devices and Internet access. Bluetooth modemis configured to communicate with another Bluetooth-capable device according to the Bluetooth short-range wireless technology standard(s) such as IEEE 802.15.1 and/or managed by the Bluetooth Special Interest Group (SIG).

1102 1182 1184 1186 1180 1180 1180 1102 1102 1104 1102 1102 1154 1152 1136 1138 1182 1102 1102 1102 1184 1102 1102 1186 1102 Computing devicecan further include power supply, LI receiver, accelerometer, and/or one or more wired interfaces. Example wired interfacesinclude a USB port, IEEE 1394 (FireWire) port, a RS-232 port, an HDMI (High-Definition Multimedia Interface) port (e.g., for connection to an external display), a DisplayPort port (e.g., for connection to an external display), an audio port, and/or an Ethernet port, the purposes and functions of each of which are well known to persons skilled in the relevant art(s). Wired interface(s)of computing deviceprovide for wired connections between computing deviceand network, or between computing deviceand one or more devices/peripherals when such devices/peripherals are external to computing device(e.g., a pointing device, display, speaker, camera, physical keyboard, etc.). Power supplyis configured to supply power to each of the components of computing deviceand receives power from a battery internal to computing device, and/or from a power cord plugged into a power port of computing device(e.g., a USB port, an A/C power port). LI receiveris useable for location determination of computing deviceand in examples includes a satellite navigation receiver such as a Global Positioning System (GPS) receiver and/or includes other type of location determiner configured to determine location of computing devicebased on received information (e.g., using cell tower triangulation, etc.). Accelerometer, when present, is configured to determine an orientation of computing device.

1102 1102 1110 1156 1102 Note that the illustrated components of computing deviceare not required or all-inclusive, and fewer or greater numbers of components can be present as would be recognized by one skilled in the art. In examples, computing deviceincludes one or more of a gyroscope, barometer, proximity sensor, ambient light sensor, digital compass, etc. In an example, processorand memoryare co-located in a same semiconductor device package, such as being included together in an integrated circuit chip, FPGA, or system-on-chip (SOC), optionally along with further components of computing device.

1102 1120 1110 In embodiments, computing deviceis configured to implement any of the above-described features of flowcharts herein. Computer program logic for performing any of the operations, steps, and/or functions described herein is stored in storageand executed by processor.

1170 1100 1102 1104 1170 1170 1172 1172 1172 1174 1174 1104 1174 1104 1174 11 FIG. 11 FIG. In some embodiments, server infrastructureis present in computing environmentand is communicatively coupled with computing devicevia network. Server infrastructure, when present, is a network-accessible server set (e.g., a cloud-based environment or platform). As shown in, server infrastructureincludes clusters. Each of clusterscomprises a group of one or more compute nodes and/or a group of one or more storage nodes. For example, as shown in, clusterincludes nodes. Each of nodesare accessible via network(e.g., in a “cloud-based” embodiment) to build, deploy, and manage applications and services. In examples, any of nodesis a storage node that comprises a plurality of physical storage disks, SSDs, and/or other physical storage devices that are accessible via networkand are configured to store data associated with the applications and services managed by nodes.

1174 1174 1102 1174 1174 1146 1148 1158 1110 1142 1144 1102 1148 1176 1178 1158 1176 1178 1146 1174 1176 11 FIG. Each of nodes, as a compute node, comprises one or more server computers, server systems, and/or computing devices. For instance, a nodein accordance with an embodiment includes one or more of the components of computing devicedisclosed herein. Each of nodesis configured to execute one or more software applications (or “applications”) and/or services and/or manage hardware resources (e.g., processors, memory, etc.), which are utilized by users (e.g., customers) of the network-accessible server set. In examples, as shown in, nodesincludes a nodethat includes storageand/or one or more of a processor(e.g., similar to processor, GPU, and/or NPUof computing device). Storagestores application programsand application data. Processor(s)operate application programswhich access and/or generate related application data. In an implementation, nodes such as nodeof nodesoperate or comprise one or more virtual machines, with each virtual machine emulating a system architecture (e.g., an operating system), in an isolated manner, upon which applications such as application programsare executed.

1172 1172 1100 In embodiments, one or more of clustersare located/co-located (e.g., housed in one or more nearby buildings with associated components such as backup power supplies, redundant data communications, environmental controls, etc.) to form a datacenter, or are arranged in other manners. Accordingly, in an embodiment, one or more of clustersare included in a datacenter in a distributed collection of datacenters. In embodiments, exemplary computing environmentcomprises part of a cloud-based platform.

1102 1176 1102 In an embodiment, computing deviceaccesses application programsfor execution in any manner, such as by a client application and/or a browser at computing device.

1102 1114 1116 1170 1176 1178 1112 1114 1120 1170 In an example, for purposes of network (e.g., cloud) backup and data security, computing deviceadditionally and/or alternatively synchronizes copies of application programsand/or application datato be stored at network-based server infrastructureas application programsand/or application data. In examples, operating systemand/or application programsinclude a file hosting service client configured to synchronize applications and/or data stored in storageat network-based server infrastructure.

1192 1100 1102 1104 1192 1192 1198 1192 1102 1192 1196 1102 1192 1194 1196 1198 1190 1110 1142 1144 1102 1196 1190 1196 1102 1114 1116 1192 1196 1198 In some embodiments, on-premises serversare present in computing environmentand are communicatively coupled with computing devicevia network. On-premises servers, when present, are hosted within an organization's infrastructure and, in many cases, physically onsite of a facility of that organization. On-premises serversare controlled, administered, and maintained by IT (Information Technology) personnel of the organization or an IT partner to the organization. Application datacan be shared by on-premises serversbetween computing devices of the organization, including computing device(when part of an organization) through a local network of the organization, and/or through further networks accessible to the organization (including the Internet). Furthermore, in examples, on-premises serversserve applications such as application programsto the computing devices of the organization, including computing device. Accordingly, in examples, on-premises serversinclude storage(which includes one or more physical storage devices such as storage disks and/or SSDs) for storage of application programsand application dataand include a processor(e.g., similar to processor, GPU, and/or NPUof computing device) for execution of application programs. In some embodiments, multiple processorsare present for execution of application programsand/or for other purposes. In further examples, computing deviceis configured to synchronize copies of application programsand/or application datafor backup storage at on-premises serversas application programsand/or application data.

1102 1170 1192 1102 1102 1170 1192 Embodiments described herein may be implemented in one or more of computing device, network-based server infrastructure, and on-premises servers. For example, in some embodiments, computing deviceis used to implement systems, clients, or devices, or components/subcomponents thereof, disclosed elsewhere herein. In other embodiments, a combination of computing device, network-based server infrastructure, and/or on-premises serversis used to implement the systems, clients, or devices, or components/subcomponents thereof, disclosed elsewhere herein.

1120 As used herein, the terms “computer program medium,” “computer-readable medium,” “computer-readable storage medium,” and “computer-readable storage device,” etc., are used to refer to physical hardware media. Examples of such physical hardware media include any hard disk, optical disk, SSD, other physical hardware media such as RAMs, ROMs, flash memory, digital video disks, zip disks, MEMs (microelectronic machine) memory, nanotechnology-based storage devices, and further types of physical/tangible hardware storage media of storage. Such computer-readable media and/or storage media are distinguished from and non-overlapping with communication media, propagating signals, and signals per se. Stated differently, “computer program medium,” “computer-readable medium,” “computer-readable storage medium,” and “computer-readable storage device” do not encompass communication media, propagating signals, and signals per se. Communication media embodies computer-readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave. The term “modulated data signal” means a signal that has one or more of its characteristics set or changed in such a manner as to encode information in the signal. By way of example, and not limitation, communication media includes wireless media such as acoustic, RF, infrared, and other wireless media, as well as wired media. Embodiments are also directed to such communication media that are separate and non-overlapping with embodiments directed to computer-readable storage media.

1114 1120 1160 1160 1104 1102 1102 As noted above, computer programs and modules (including application programs) are stored in storage. Such computer programs can also be received via wired interface(s)and/or wireless modem(s)over network. Such computer programs, when executed or loaded by an application, enable computing deviceto implement features of embodiments discussed herein. Accordingly, such computer programs represent controllers of the computing device.

1120 Embodiments are also directed to computer program products comprising computer code or instructions stored on any computer-readable medium or computer-readable storage medium. Such computer program products include the physical storage of storageas well as further physical storage types.

In embodiments, a system comprises: a processor; and a memory device comprising program code executable to cause the processor to: receive event data associated with a set of objects; determine, based on rule hints determined from a set of event trigger rules, a set of logical partitions for partitioning the event data, a first logical partition of the set of logical partitions comprising first event data associated with a first object of the set of objects; provide, to a first event processor of a set of event processors, the first logical partition to enable the first event processor to update a state associated with the first object based on the first event data in the first logical partition; determine that the state associated with the first object satisfies a first trigger condition specified by a first event trigger rule of the set of event trigger rules; and perform a first action specified by the first event trigger rule.

In embodiments, to determine the set of logical partitions for partitioning the event data, the program code is executable to cause the processor to: determine a set of capacities of the set of event processors; and determine the set of logical partitions based on the set of capacities of the set of event processors.

In embodiments, the program code is executable to further cause the processor to: determine a set of updated capacities of the set of event processors; determine, based on the set of updated capacities, a set of updated logical partitions for repartitioning the event data, a first updated logical partition of the set of updated logical partitions comprising second event data associated with the first object; access, by a second event processor of the set of event processors, state information associated with the first object; and provide, to the second event processor, the first updated logical partition to enable the second event processor to update the state associated with the first object based on the second event data in the first updated logical partition.

In embodiments, the program code is executable to further cause the processor to: periodically store, by the first event processor, a progress indicator indicative of the first event data processed by the first event processor; transfer processing of the first logical partition from the first event processor to a second event processor of the set of event processors; access, by the second event processor, the first process indicator and state information associated with the first object; redirect the first logical partition from the first event processor to the second event processor; and process, by the second event processor, the first event data in the first logical partition starting from a position determined based on the progress indicator.

In embodiments, the program code is executable to further cause the processor to: add a new event trigger rule to the set of event trigger rules, the new event trigger rule associated with a second object of the set of objects; replay event data associated with the second object; update a state associated with the second object based on the event data associated with the second object; determine that the state associated with the second object satisfies a second trigger condition specified by the new event trigger rule; and perform a second action specified by the new event trigger rule.

In embodiments, the program code is executable to further cause the processor to: determine, prior to partitioning the event data, that the event data satisfies a second trigger condition specified by a stateless event trigger rule; and perform a second action specified by the stateless event trigger rule.

In embodiments, to perform the first action, the program code is executable to cause the processor to perform at least one of: transmit a message to a recipient associated with the first object; open, in a ticketing system, a ticket for the first object; invoke a webhook associated with the first object; call a function of an application program interface (API); or automatically perform a remedial action to alter the state of the first object.

In embodiments, a method comprises: receiving event data associated with a set of objects; determining, based on rule hints determined from a set of event trigger rules, a set of logical partitions for partitioning the event data, a first logical partition of the set of logical partitions comprising first event data associated with a first object of the set of objects; providing, to a first event processor of a set of event processors, the first logical partition to enable the first event processor to update a state associated with the first object based on the first event data in the first logical partition; determining that the state associated with the first object satisfies a first trigger condition specified by a first event trigger rule of the set of event trigger rules; and performing a first action specified by the first event trigger rule.

In embodiments, determining, based on rule hints determined from a set of event trigger rules, a set of logical partitions for partitioning the event data comprises: determining a set of capacities of the set of event processors; and determining the set of logical partitions based on the set of capacities of the set of event processors.

In embodiments, the method further comprises: determining a set of updated capacities of the set of event processors; determining, based on the set of updated capacities, a set of updated logical partitions for repartitioning the event data, a first updated logical partition of the set of updated logical partitions comprising second event data associated with the first object; accessing, by a second event processor of the set of event processors, state information associated with the first object; and providing, to the second event processor, the first updated logical partition to enable the second event processor to update the state associated with the first object based on the second event data in the first updated logical partition.

In embodiments, the method further comprises: periodically storing, by the first event processor, a progress indicator indicative of the first event data processed by the first event processor; transferring processing of the first logical partition from the first event processor to a second event processor of the set of event processors; accessing, by the second event processor, the first process indicator and state information associated with the first object; redirecting the first logical partition from the first event processor to the second event processor; and processing, by the second event processor, the first event data in the first logical partition starting from a position determined based on the progress indicator.

In embodiments, the method further comprises: adding a new event trigger rule to the set of event trigger rules, the new event trigger rule associated with a second object of the set of objects; replaying event data associated with the second object; updating a state associated with the second object based on the event data associated with the second object; determining that the state associated with the second object satisfies a second trigger condition specified by the new event trigger rule; and performing a second action specified by the new event trigger rule.

In embodiments, the method further comprises: determining, prior to partitioning the event data, that the event data satisfies a second trigger condition specified by a stateless event trigger rule; and performing a second action specified by the stateless event trigger rule.

In embodiments, performing a first action comprises at least one of: transmitting a message to a recipient associated with the first object; opening, in a ticketing system, a ticket for the first object; invoking a webhook associated with the first object; calling a function of an application program interface (API); or automatically performing a remedial action to alter the state of the first object.

In embodiments, a computer-readable storage medium comprises executable instructions that, when executed by a processor, cause the processor to: receive event data associated with a set of objects; determine, based on rule hints determined from a set of event trigger rules, a set of logical partitions for partitioning the event data, a first logical partition of the set of logical partitions comprising first event data associated with a first object of the set of objects; provide, to a first event processor of a set of event processors, the first logical partition to enable the first event processor to update a state associated with the first object based on the first event data in the first logical partition; determine that the state associated with the first object satisfies a first trigger condition specified by a first event trigger rule of the set of event trigger rules; and perform a first action specified by the first event trigger rule.

In embodiments, to determine the set of logical partitions for partitioning the event data, the executable instructions, when executed by the processor, cause the processor to: determine a set of capacities of the set of event processors; and determine the set of logical partitions based on the set of capacities of the set of event processors.

In embodiments, the executable instructions, when executed by the processor, further cause the processor to: determine a set of updated capacities of the set of event processors; determine, based on the set of updated capacities, a set of updated logical partitions for repartitioning the event data, a first updated logical partition of the set of updated logical partitions comprising second event data associated with the first object; access, by a second event processor of the set of event processors, state information associated with the first object; and provide, to the second event processor, the first updated logical partition to enable the second event processor to update the state associated with the first object based on the second event data in the first updated logical partition.

In embodiments, the executable instructions, when executed by the processor, further cause the processor to: periodically store, by the first event processor, a progress indicator indicative of the first event data processed by the first event processor; transfer processing of the first logical partition from the first event processor to a second event processor of the set of event processors; access, by the second event processor, the first process indicator and state information associated with the first object; redirect the first logical partition from the first event processor to the second event processor; and process, by the second event processor, the first event data in the first logical partition starting from a position determined based on the progress indicator.

In embodiments, the executable instructions, when executed by the processor, further cause the processor to: add a new event trigger rule to the set of event trigger rules, the new event trigger rule associated with a second object of the set of objects; replay event data associated with the second object; update a state associated with the second object based on the event data associated with the second object; determine that the state associated with the second object satisfies a second trigger condition specified by the new event trigger rule; and perform a second action specified by the new event trigger rule.

In embodiments, the executable instructions, when executed by the processor, further cause the processor to: determine, prior to partitioning the event data, that the event data satisfies a second trigger condition specified by a stateless event trigger rule; and perform a second action specified by the stateless event trigger rule.

References in the specification to “one embodiment,” “an embodiment,” “an example embodiment,” etc., indicate that the embodiment described may include a particular feature, structure, or characteristic, but every embodiment may not necessarily include the particular feature, structure, or characteristic. Moreover, such phrases are not necessarily referring to the same embodiment. Further, when a particular feature, structure, or characteristic is described in connection with an embodiment, it is submitted that it is within the knowledge of one skilled in the art to effect such feature, structure, or characteristic in connection with other embodiments whether or not explicitly described.

In the discussion, unless otherwise stated, adjectives such as “substantially” and “about” modifying a condition or relationship characteristic of a feature or features of an embodiment of the disclosure, are understood to mean that the condition or characteristic is defined to within tolerances that are acceptable for operation of the embodiment for an application for which it is intended. Furthermore, where “based on” is used to indicate an effect being a result of an indicated cause, it is to be understood that the effect is not required to only result from the indicated cause, but that any number of possible additional causes may also contribute to the effect. Thus, as used herein, the term “based on” should be understood to be equivalent to the term “based at least on.”

While various embodiments of the present disclosure have been described above, it should be understood that they have been presented by way of example only, and not limitation. It will be understood by those skilled in the relevant art(s) that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined in the appended claims. Accordingly, the breadth and scope of the present invention should not be limited by any of the above-described exemplary embodiments, but should be defined only in accordance with the following claims and their equivalents.